神经网络在柴油机故障诊断中的应用研究  

The Application Research of Neural Networks to the Fault Diagnosis of Diesel Engine

作　　者：王悦民[1] 魏强 潜伟建[1] 

机构地区：[1]武汉海军工程大学内燃动力系,湖北武汉430033 [2]武汉七零一研究所,湖北武汉430064

出　　处：《系统工程理论与实践》2001年第10期52-57,共6页Systems Engineering-Theory & Practice

摘　　要：该文利用尺度共轭梯度算法的神经网络对柴油机进行故障诊断 .首先讨论了其训练算法 ,然后确定了柴油机故障诊断所用特征参数及故障种类 ,并提出特征参数数据归一化公式 ,最后以 6-1 3 5 ZC柴油机为例 ,将实验数据输入网络验证 .结果表明 ,神经网络对柴油机故障识别率很高 ,应用于柴油机故障诊断领域是切实可行的 .This paper applies neural networks to diagnose the fault of diesel engines by using Scaled Conjugate Gradient Algorithm. First, the Scaled Conjugate Gradient Algorithm is discussed, then,the feature parameters and the kinds of fault used by diagnosing diesel engine fault are given. In addition, the best fit curve equations of feature parameters are extracted. At last, 6-135ZC diesel engine is exampled for this, and the neural networks is verified by using the fault data of diesel engine . Experimental result demonstrates that the result of diagnosis to diesel engine fault is of great accuracy , the application of neural networks to this field is useful and reliable .

关 键 词：柴油机 神经网络 故障诊断 共轭梯度法 

分 类 号：TK438[动力工程及工程热物理—动力机械及工程] TP183[自动化与计算机技术—控制理论与控制工程]